Anything captured in one client is available in the next, which is the practical answer to re-explaining yourself every time you switch tools. Each account is isolated at the database level by row-level security, and the published AI Use Policy states the data is not used to train models.
A hosted MCP server that carries a knowledge layer about you — your Core Imprint, Knowledge Vault, Instincts and captured patterns — into every AI tool you connect. Where a company knowledge base brings company knowledge to AI, this is built around the individual operator, consultant or founder.
- Core Imprint access — connected clients read your identity and adapt style to it
- Knowledge search — query your stored knowledge base from any connected tool
- Knowledge capture — tell the assistant to remember something and it is stored for later sessions
- Task management — create, update and track tasks in conversation
- File storage — save and retrieve documents across sessions
- Network intelligence — query stored contact and relationship data
- Learning — session patterns feed the Tempering system over time
- Higher tiers add Agents (AI personalities), Chains (multi-step orchestration) and Decision Gates (behavioural rules)
A Tempreon account, and about 15 minutes for the 13-dimension onboarding that seeds your Core Imprint. Connection is a single URL — https://api.tempreon.com/functions/v1/tempreon-mcp/mcp, and the `/mcp` suffix is the most common thing people leave off. Transport is Streamable HTTP, so there is no local config file and no process to manage. Authentication is OAuth 2.1 with PKCE and dynamic client registration per RFC 7591, so no pre-registration and no key to paste; protected-resource metadata is discoverable at `/.well-known/oauth-protected-resource`. Setup guides ship for Claude Desktop, Claude Code, Codex, ChatGPT and Cursor. Free and paid plans, paid from $19/month.